No, but correlation isn't meaningless either. This study [1] appears to be as robust as one can get without doing a double-blind trial. Given the effect size, I suspect that is in the works.
Correlation is sufficient to draw additional studies to add to the discussion however.
And with a proper survey, you can actually account for a lot of these effects. (Ex: after accounting for outdoor temperatures, the Ice Cream Sales / Drowning correlation can be largely factored out).